CN115980530A - High-voltage cable partial discharge mode identification method, system, equipment and storage medium - Google Patents

Abstract

The invention discloses a method, a system, equipment and a storage medium for identifying a partial discharge mode of a high-voltage cable, wherein the method comprises the following steps: acquiring partial discharge data of a high-voltage cable, extracting partial discharge transient pulses from the partial discharge data of the high-voltage cable, determining characteristic parameters of the partial discharge transient pulses, and performing normalization processing on the characteristic parameters; and inputting the result after the normalization processing into an optimized hierarchical pattern recognition system to recognize and obtain the partial discharge pattern of the high-voltage cable.

Description

High-voltage cable partial discharge mode identification method, system, equipment and storage medium

Technical Field

The invention belongs to the field of power equipment, and relates to a method, a system, equipment and a storage medium for identifying a partial discharge mode of a high-voltage cable.

Background

High voltage cables are used more and more widely as important components in power systems. The partial discharge pattern recognition of a high-voltage cable is one of the important links for monitoring and diagnosing the state of the high-voltage cable. However, there are many kinds of partial discharges, such as free particle discharge, floating potential discharge, creeping discharge, etc., and the similarity between partial discharge types is large, so that a pattern recognition method with high accuracy is required. To solve the above problems, more and more methods are applied to the field of partial discharge pattern recognition.

At present, the machine learning method is mainly used for identifying the partial discharge mode of the high-voltage cable, and the method can be divided into two types: the "white box" and "black box" methods. "black box" algorithms such as CNN, SDAE, etc., but this method requires a large amount of training data and does not visually reveal the relationship between different cable defect types and rules. The white box algorithm has stronger interpretability and is easier to realize in on-line monitoring programming.

Decision Trees (DTs) are the most widely used white-box algorithms in pattern recognition, however, the recognition capability of a single tree is limited, and the problem of overfitting is easily caused, which in turn leads to poor recognition accuracy.

Disclosure of Invention

The present invention is directed to overcome the above drawbacks of the prior art, and provides a method, a system, a device and a storage medium for identifying a partial discharge mode of a high voltage cable, which can identify the partial discharge mode of the high voltage cable more accurately.

In order to achieve the purpose, the invention adopts the following technical scheme:

in one aspect, the present invention provides a method for identifying a partial discharge mode of a high voltage cable, including:

acquiring partial discharge data of a high-voltage cable, extracting partial discharge transient pulses from the partial discharge data of the high-voltage cable, determining characteristic parameters of the partial discharge transient pulses, and performing normalization processing on the characteristic parameters;

and inputting the result after the normalization processing into an optimized layered pattern recognition system so as to recognize and obtain the partial discharge pattern of the high-voltage cable.

Before the inputting the result after the normalization processing into the optimized hierarchical pattern recognition system, the method further comprises:

acquiring partial discharge data samples of different high-voltage cable insulation defect types, extracting partial discharge instantaneous pulse samples from the partial discharge data samples, constructing characteristic parameters of the partial discharge instantaneous pulse samples, performing normalization processing to obtain original characteristic data, and dividing the original characteristic data into a training set and a test set;

establishing a layered pattern recognition system;

and optimizing the hierarchical pattern recognition system based on the training set and the test set to obtain the optimized hierarchical pattern recognition system.

The hierarchical pattern recognition system comprises a first layer which comprises a random forest model, a gradient boosting decision tree model and a first regularized greedy forest model, a second layer which comprises a second regularized greedy forest model, and a probability value obtained by predicting the random forest model (RF) in the first layer, a probability value obtained by predicting the gradient boosting decision tree model (GBDT) and a probability value obtained by predicting the first regularized greedy forest model (RGF) are used as input of the second regularized greedy forest model in the second layer.

The specific process of optimizing the hierarchical pattern recognition system based on the training set and the test set is as follows:

and optimizing the hierarchical pattern recognition system by using a Harris eagle algorithm based on the training set and the testing set.

In a second aspect of the present invention, the present invention provides a system for identifying a partial discharge mode of a high voltage cable, comprising:

the device comprises a first acquisition module, a second acquisition module and a control module, wherein the first acquisition module is used for acquiring partial discharge data of a high-voltage cable, extracting a partial discharge transient pulse from the partial discharge data of the high-voltage cable, determining a characteristic parameter of the partial discharge transient pulse and carrying out normalization processing on the characteristic parameter;

and the identification module is used for inputting the result after the normalization processing into the optimized layered pattern identification system so as to identify and obtain the partial discharge pattern of the high-voltage cable.

Further comprising:

the second acquisition module is used for acquiring partial discharge data samples of different high-voltage cable insulation defect types, extracting partial discharge instantaneous pulse samples from the partial discharge data samples, constructing characteristic parameters of the partial discharge instantaneous pulse samples, performing normalization processing to obtain original characteristic data, and dividing the original characteristic data into a training set and a test set;

the establishing module is used for establishing a layered pattern recognition system;

and the optimization module is used for optimizing the hierarchical pattern recognition system based on the training set and the testing set to obtain the optimized hierarchical pattern recognition system.

The hierarchical pattern recognition system comprises a first layer and a second layer, wherein the first layer comprises a random forest model, a gradient boosting decision tree model and a first regularized greedy forest model, the second layer comprises a second regularized greedy forest model, and a probability value obtained by predicting a random forest model (RF) in the first layer, a probability value obtained by predicting a gradient boosting decision tree model (GBDT) and a probability value obtained by predicting a first regularized greedy forest model (RGF) are used as input of the second regularized greedy forest model in the second layer.

And optimizing the hierarchical pattern recognition system by using a Harris eagle algorithm based on the training set and the test set.

In one aspect, the present invention provides a computer device, which includes a memory, a processor and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the high voltage cable partial discharge pattern recognition method when executing the computer program.

In a fourth aspect of the present invention, the present invention provides a computer-readable storage medium storing a computer program, which when executed by a processor, implements the steps of the high voltage cable partial discharge pattern recognition method.

The invention has the following beneficial effects:

when the method, the system, the equipment and the storage medium for identifying the partial discharge mode of the high-voltage cable are specifically operated, the layered mode identification system is applied to the identification of the partial discharge mode of the high-voltage cable, the accuracy of identification is improved through model fusion, specifically, the result after normalization processing is input into the optimized layered mode identification system to identify and obtain the partial discharge mode of the high-voltage cable, and the method, the system and the equipment are simple and convenient to operate and convenient to implement.

Further, the hierarchical pattern recognition system is optimized based on the harris eagle algorithm to reduce errors of model prediction.

Drawings

The accompanying drawings, which are incorporated in and constitute a part of this specification, are included to provide a further understanding of the invention, and are incorporated in and constitute a part of this specification, illustrate exemplary embodiments of the invention and together with the description serve to explain the invention and not to limit the invention. In the drawings:

FIG. 1 is a flow chart of the present invention;

FIG. 2 is a block diagram of a two-tier identification system;

FIG. 3 is a flow chart of Harris eagle algorithm optimizing hierarchical identification system parameters.

Detailed Description

In order to make those skilled in the art better understand the technical solutions of the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.

It should be noted that the terms "first," "second," and the like in the description and claims of the present invention and in the drawings described above are used for distinguishing between similar elements and not necessarily for describing a particular sequential or chronological order. It is to be understood that the data so used is interchangeable under appropriate circumstances such that the embodiments of the invention described herein are capable of operation in other sequences than those illustrated or described herein. Furthermore, the terms "comprises," "comprising," and "having," and any variations thereof, are intended to cover a non-exclusive inclusion, such that a process, method, system, article, or apparatus that comprises a list of steps or elements is not necessarily limited to those steps or elements expressly listed, but may include other steps or elements not expressly listed or inherent to such process, method, article, or apparatus.

The invention is described in further detail below with reference to the accompanying drawings:

example one

Referring to fig. 1 and 2, the method for identifying the partial discharge mode of the high-voltage cable based on multi-model fusion according to the present invention includes the following steps:

1) Acquiring partial discharge data of different high-voltage cable insulation defect types, extracting partial discharge transient pulses from the partial discharge data, constructing characteristic parameters of the partial discharge transient pulses, carrying out normalization processing on the characteristic parameters to obtain original characteristic data, and dividing the original characteristic data into a training set and a test set;

2) Constructing a hierarchical pattern recognition system, wherein a first layer of the hierarchical pattern recognition system is a random forest model (RF), a gradient boosting decision tree model (GBDT) and a regularized greedy forest model (RGF), and a second layer of the hierarchical pattern recognition system is the regularized greedy forest model, wherein a probability value obtained by predicting the random forest model (RF) in the first layer, a probability value obtained by predicting the gradient boosting decision tree model (GBDT) and a probability value obtained by predicting the first regularized greedy forest model (RGF) are used as input of a second regularized greedy forest model in the second layer to form the hierarchical pattern recognition system;

3) Optimizing parameters of the hierarchical recognition system by using a Harris eagle algorithm based on a training set and a test set;

4) And identifying the partial discharge mode of the high-voltage cable by using the optimized layered identification system.

It should be noted that the Random Forest (RF) is a collective learning method, and is composed of many single classification trees (CART). And the RF adopts a Bootstrap resampling method to extract a plurality of groups of samples from the sample set as training sets, carries out CART decision tree modeling on each training set respectively, and votes through each decision tree to obtain a final classification result. The performance of the RF depends on the classification accuracy of each decision tree and the correlation between the decision trees. The greater the classification accuracy of the decision trees and the smaller the correlation between decision trees, the better the performance of the RF. Therefore, increasing the classification strength of each decision tree and reducing the correlation between trees can effectively reduce the overall error rate of RF. The RF construction process comprises the following steps:

11 Using a Bootstrap method to extract k training sample sets with the same number of samples from an original sample set S;

12 For each generated training sample set, utilizing CART algorithm to learn and generate k decision trees, wherein the segmentation features of each node of CART are selected from a sub-feature set (the sub-feature set is composed of F features randomly extracted from all the features);

3) The k CART decision trees are combined to form the RF model.

The Gradient boost decision tree model (GBDT) is formed by integrating decision trees in a Gradient Boosting mode, so that the GBDT has high classification precision and strong generalization capability. On the basis of accurately processing the two-classification problem, the GBDT can be applied to a multi-classification scene through combination, and the construction of the two-classification GBDT is as follows:

1) Setting iteration number M, setting current iteration number M to be 0, and initializing a first classifier as shown in formula (1):

2) Calculating a negative gradient value of a current loss function

Namely that

For the classification problem, the loss function is a logarithmic loss of:

L(y _i ,F(x _i ))＝-(y _i logp _i +(1-y _i )log(1-p _i )) (3)

wherein the content of the first and second substances,

3) To pair

Fitting is carried out, and the parameters of each leaf node are calculated

Wherein R is _jm And classifying the tree for the J leaf node at the mth iteration.

4) And updating the classifier model according to the leaf node parameters and the learning rate eta.

5) And (5) repeating the steps 2) to 4) for M times, adding 1 to the iteration number M each time, and generating a final binary GBDT model when the iteration is finished.

The regularized greedy forest model (RGF) is improved on the basis of GBDT, and the core idea is as follows: every iteration is not only optimized for a newly-built tree, but the whole greedy forest is learned, global parameters are optimized after a decision tree is newly added, and an explicit regular function is added to prevent overfitting.

In a decision tree, a path formed from a root node x of the tree to a child node v of the tree is a classification rule, and the classification rule is formulated as:

wherein x [ i ] _j ]A j-th characteristic value representing an i-th sample; t is t _j A threshold value for the jth feature; i (-) indicates that the true result in parentheses is 1, otherwise 0. In a single decision tree, contain v ₁ ，v ₂ The node v of the two children nodes can be denoted b _v (x)＝b _v1 (x)+b _v2 (x) In that respect The decision tree model can be seen as a combined model of the leaf nodes, i.e.:

wherein, a _v Representing the weight parameters of the nodes and F representing the decision forest.

When RGF constructs decision forest, a loss function Q = L (h) is set _F (x),Y)+R(h _F ) Wherein R (h) _F ) As a regularization term, L (h) _F (x) Y) is defined according to the problem to be treated, L (h) in the present invention _F (x) And Y) is:

wherein Y represents a sample label.

As shown in FIG. 3, the present invention utilizes the Harris eagle algorithm (HHO) to optimize model parameters for a hierarchical recognition system. HHO is a bionic algorithm for simulating the predation behavior of the hawk and mainly comprises 3 parts, namely an exploration phase, a transition phase from the exploration to the development and a development phase.

Stage 1: exploration phase

At this stage, harris eagle is in a wait state, performing a prey search with two strategies:

wherein X (t) refers to the current position of eagle, X _rand (t) randomly selecting individuals from the current eagle population; x _rabbit (t) is prey position; x _m (t) is the average position of the current population, r ₁ 、r ₂ 、r ₃ 、r ₄ And q is a random number between 0 and 1; ub and lb are the upper and lower bounds of the search space; n is the total number of the hawks, namely the population number.

And (2) stage: transition phase

Harris hawk switches between different development activities depending on the escape energy factor E of the prey, which is:

wherein E is the escape energy of the prey; e ₀ Is the initial energy of the prey; t is the current iteration number; and T is the maximum iteration number.

And (3) stage: development phase

Aiming at the actual hunting process, harris hawk utilizes four strategies to carry out the position updating of the development stage so as to better simulate the hunting behavior.

3.1 Soft surround

When | E | ≧ 0.5 and r ≧ 0.5, the prey has enough energy to begin trying to escape from the enclosure through random walk, but finally cannot escape, so Harris hawk adopts a soft enclosure scheme for hunting, and the mathematical expression is as follows:

X(t+1)＝X _rabbit (t)-X(t)-E|JX _rabbit (t)-X(t)| (12)

wherein J is a random number between 0 and 2.

3.2 hard surround

When | E | <0.5 and r ≧ 0.5, the prey does not have enough energy to escape and there is no chance of escape, so Harris eagle uses the hard surrounding scheme for hunting, the mathematical expression is:

X(t+1)＝X _rabbit (t)-E|X _rabbit (t)-X(t)| (13)

3.3 gradual quick dive soft surround

When E | ≧ 0.5 and r <0.5, the prey is likely to escape from the enclosure and the escape energy is enough, harris hawk will form an intelligent soft enclosure during attack, and implement two strategies, when the first strategy fails, the second strategy is adopted, the mathematical expression is

Y＝X _rabbit (t)-E|JX _rabbit (t)-X(t)| (15)

Z＝Y+S×LF(D) (16)

Wherein D is the dimension involved in the objective function; s is a D-dimensional random vector; LF is the Levy flight function.

3.4 progressive fast dive hard surround

When | E | <0.5 and r <0.5, there is a possibility that the prey will escape, but the escape energy is insufficient, harris hawks will form an intelligent hard enclosure during the attack to reduce the average distance to the prey, and the phase update strategy is

Y＝X _rabbit (t)-E|JX _rabbit (t)-X _m (t)| (18)

Z＝Y+S×LF(D) (19)。

Example two

The invention relates to a high-voltage cable partial discharge mode identification system, which comprises:

the device comprises a first acquisition module, a second acquisition module and a control module, wherein the first acquisition module is used for acquiring partial discharge data of a high-voltage cable, extracting a partial discharge transient pulse from the partial discharge data of the high-voltage cable, determining a characteristic parameter of the partial discharge transient pulse and carrying out normalization processing on the characteristic parameter;

and the identification module is used for inputting the result after the normalization processing into the optimized layered pattern identification system so as to identify and obtain the partial discharge pattern of the high-voltage cable.

Further comprising:

the second acquisition module is used for acquiring partial discharge data samples of different high-voltage cable insulation defect types, extracting partial discharge instantaneous pulse samples from the partial discharge data samples, constructing characteristic parameters of the partial discharge instantaneous pulse samples, performing normalization processing to obtain original characteristic data, and dividing the original characteristic data into a training set and a test set;

the establishing module is used for establishing a layered pattern recognition system;

and the optimization module is used for optimizing the hierarchical pattern recognition system based on the training set and the test set to obtain the optimized hierarchical pattern recognition system.

The hierarchical pattern recognition system comprises a first layer and a second layer, wherein the first layer comprises a random forest model, a gradient boosting decision tree model and a first regularized greedy forest model, the second layer comprises a second regularized greedy forest model, and a probability value obtained by predicting a random forest model (RF) in the first layer, a probability value obtained by predicting a gradient boosting decision tree model (GBDT) and a probability value obtained by predicting a first regularized greedy forest model (RGF) are used as input of the second regularized greedy forest model in the second layer.

And optimizing the hierarchical pattern recognition system by using a Harris eagle algorithm based on the training set and the test set.

EXAMPLE III

A computer device, comprising a memory, a processor and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the high voltage cable partial discharge pattern recognition method when executing the computer program, wherein the memory may comprise a memory, such as a high speed random access memory, and may further comprise a non-volatile memory, such as at least one disk memory, etc.; the processor, the network interface and the memory are connected with each other through an internal bus, wherein the internal bus can be an industrial standard system structure bus, a peripheral component interconnection standard bus, an extended industrial standard structure bus and the like, and the bus can be divided into an address bus, a data bus, a control bus and the like. The memory is used for storing programs, and particularly, the programs can include program codes which comprise computer operation instructions. The memory may include both memory and non-volatile storage and provides instructions and data to the processor.

Example four

A computer-readable storage medium, in which a computer program is stored which, when being executed by a processor, carries out the steps of the high voltage cable partial discharge pattern recognition method, in particular, but not exclusively, volatile memory and/or non-volatile memory, for example. The volatile memory may include Random Access Memory (RAM) and/or cache memory (cache), among others. The non-volatile memory may include a Read Only Memory (ROM), hard disk, flash memory, optical disk, magnetic disk, and the like.

As will be appreciated by one skilled in the art, embodiments of the present application may be provided as a method, system, or computer program product. Accordingly, the present application may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present application may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein.

The present application is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the application. It will be understood that each flow and/or block of the flow diagrams and/or block diagrams, and combinations of flows and/or blocks in the flow diagrams and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.

These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.

These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.

Finally, it should be noted that: the above embodiments are only for illustrating the technical solutions of the present invention and not for limiting the same, and although the present invention is described in detail with reference to the above embodiments, those of ordinary skill in the art should understand that: modifications and equivalents may be made to the embodiments of the invention without departing from the spirit and scope of the invention, which is to be covered by the claims.

Claims (10)

1. A method for identifying partial discharge modes of a high-voltage cable is characterized by comprising the following steps:

acquiring partial discharge data of a high-voltage cable, extracting partial discharge transient pulses from the partial discharge data of the high-voltage cable, determining characteristic parameters of the partial discharge transient pulses, and performing normalization processing on the characteristic parameters;

and inputting the result after the normalization processing into an optimized layered pattern recognition system so as to recognize and obtain the partial discharge pattern of the high-voltage cable.

2. The method for identifying the partial discharge pattern of the high-voltage cable according to claim 1, wherein before inputting the result of the normalization process into the optimized hierarchical pattern identification system, the method further comprises:

acquiring partial discharge data samples of different high-voltage cable insulation defect types, extracting partial discharge instantaneous pulse samples from the partial discharge data samples, constructing characteristic parameters of the partial discharge instantaneous pulse samples, performing normalization processing to obtain original characteristic data, and dividing the original characteristic data into a training set and a test set;

establishing a layered pattern recognition system;

and optimizing the layered pattern recognition system based on the training set and the test set to obtain the optimized layered pattern recognition system.

3. The high-voltage cable partial discharge pattern recognition method as recited in claim 2, wherein a first layer of the hierarchical pattern recognition system comprises a random forest model, a gradient boosting decision tree model and a first regularized greedy forest model, a second layer of the hierarchical pattern recognition system comprises a second regularized greedy forest model, and a probability value predicted by the random forest model (RF) in the first layer, a probability value predicted by the gradient boosting decision tree model (GBDT) and a probability value predicted by the first regularized greedy forest model (RGF) are used as inputs of the second regularized greedy forest model in the second layer.

4. The method for identifying the partial discharge pattern of the high-voltage cable according to claim 2, wherein the specific process for optimizing the hierarchical pattern recognition system based on the training set and the test set comprises:

and optimizing the hierarchical pattern recognition system by using a Harris eagle algorithm based on the training set and the testing set.

5. A high voltage cable partial discharge pattern recognition system, comprising:

the device comprises a first acquisition module, a second acquisition module and a control module, wherein the first acquisition module is used for acquiring partial discharge data of a high-voltage cable, extracting a partial discharge transient pulse from the partial discharge data of the high-voltage cable, determining a characteristic parameter of the partial discharge transient pulse and carrying out normalization processing on the characteristic parameter;

and the identification module is used for inputting the result after the normalization processing into the optimized layered pattern identification system so as to identify and obtain the partial discharge pattern of the high-voltage cable.

6. The system for identifying partial discharge patterns of a high voltage cable according to claim 5, further comprising:

the second acquisition module is used for acquiring partial discharge data samples of different high-voltage cable insulation defect types, extracting partial discharge instantaneous pulse samples from the partial discharge data samples, constructing characteristic parameters of the partial discharge instantaneous pulse samples, performing normalization processing to obtain original characteristic data, and dividing the original characteristic data into a training set and a test set;

the establishing module is used for establishing a layered pattern recognition system;

and the optimization module is used for optimizing the hierarchical pattern recognition system based on the training set and the test set to obtain the optimized hierarchical pattern recognition system.

7. The high-voltage cable partial discharge pattern recognition system as claimed in claim 6, wherein a first layer of the hierarchical pattern recognition system comprises a random forest model, a gradient boosting decision tree model and a first regularized greedy forest model, a second layer of the hierarchical pattern recognition system comprises a second regularized greedy forest model, and a probability value predicted by the random forest model (RF) in the first layer, a probability value predicted by the gradient boosting decision tree model (GBDT) and a probability value predicted by the first regularized greedy forest model (RGF) are used as inputs of the second regularized greedy forest model in the second layer.

8. The system of claim 6, wherein the hierarchical pattern recognition system is optimized using a harris eagle algorithm based on the training set and the testing set.

9. Computer arrangement comprising a memory, a processor and a computer program stored in the memory and executable on the processor, characterized in that the processor realizes the steps of the method for identifying partial discharge patterns of a high voltage cable according to any of claims 1 to 4 when executing the computer program.

10. A computer-readable storage medium, in which a computer program is stored, which, when being executed by a processor, carries out the steps of the method for identifying partial discharge patterns of a high voltage cable according to any one of claims 1 to 4.

Images